ORIGINAL PAPER

Maternal Thimerosal Exposure Results in Aberrant CerebellarOxidative Stress, Thyroid Hormone Metabolism, and MotorBehavior in Rat Pups; Sex- and Strain-Dependent Effects

Z. L. Sulkowski & T. Chen & S. Midha & A. M. Zavacki &Elizabeth M. Sajdel-Sulkowska

Published online: 21 October 2011# Springer Science+Business Media, LLC 2011

Abstract Methylmercury (Met-Hg) and ethylmercury (Et-Hg) are powerful toxicants with a range of harmful neurolog-ical effects in humans and animals. While Met-Hg is arecognized trigger of oxidative stress and an endocrinedisruptor impacting neurodevelopment, the developmentalneurotoxicity of Et-Hg, a metabolite of thimerosal (TM), hasnot been explored. We hypothesized that TM exposure duringthe perinatal period impairs central nervous system develop-ment, and specifically the cerebellum, by the mechanisminvolving oxidative stress. To test this, spontaneously hyper-tensive rats (SHR) or Sprague–Dawley (SD) rat dams wereexposed to TM (200 μg/kg body weight) during pregnancy(G10–G15) and lactation (P5–P10).Male and female neonateswere evaluated for auditory and motor function; cerebellawere analyzed for oxidative stress and thyroid metabolism.TM exposure resulted in a delayed startle response in SDneonates and decreased motor learning in SHR male (22.6%),in SD male (29.8%), and in SD female (55.0%) neonates. TMexposure also resulted in a significant increase in cerebellarlevels of the oxidative stress marker 3-nitrotyrosine in SHRfemale (35.1%) and SD male (14.0%) neonates. The activity

of cerebellar type 2 deiodinase, responsible for local intra-brain conversion of thyroxine to the active hormone, 3′,3,5-triiodothyronine (T3), was significantly decreased in TM-exposed SHR male (60.9%) pups. This coincided with anincreased (47.0%) expression of a gene negatively regulatedby T3, Odf4 suggesting local intracerebellar T3 deficiency.Our data thus demonstrate a negative neurodevelopmentalimpact of perinatal TM exposure which appears to be bothstrain- and sex-dependent.

Keywords Ethylmercury . Rat . Cerebellum . Oxidativestress marker 3-nitrotrosine (3-NT) . Type 2 deiodinase (D2)

Introduction

Environmental toxicants such as heavy metals [1] includingmercury Hg [2, 3] have been identified as factors exerting arange of harmful neurological and cognitive effects inhumans and experimental animals, and have been implicatedin the etiology of a number of neuropsychiatric disorders.The major environmental organic compounds of mercuryinclude methylmercury (Met-Hg) and ethylmercury (Et-Hg).The main exposure to Met-Hg comes from contaminated fishthrough bioaccumulation of both organic and inorganic ofHg environmental contamination.

Met-Hg accumulates in both fetal and neonatal brainspotentially affecting neurodevelopment [4]. Met-Hg hasbeen shown to cross the placenta [5] and can be transferredfrom plasma to mothers’ milk [6]. It is a known trigger ofoxidative stress [7, 8] and both an endocrine [9, 10] andantioxidant defense system [11, 12] disruptor. Gestationalexposure to Met-Hg in mice results in increased lipidperoxidation and reduced developmental increase in GSHin the brain [13].

Z. L. Sulkowski : S. Midha : E. M. Sajdel-SulkowskaDepartment of Psychiatry,Harvard Medical School and Brigham and Women’s Hospital,Boston, MA, USA

T. Chen :A. M. ZavackiThyroid Section, Division of Endocrinology,Diabetes and Hypertension, Department of Medicine,Harvard Medical School and Brigham and Women’s Hospital,Boston, MA, USA

E. M. Sajdel-Sulkowska (*)Department of Psychiatry BWH, Harvard Institute of Medicine,Rm. 921, 77 Avenue Louis Pasteur,Boston, MA 02115, USAe-mail: esulkowska@rics.bwh.harvard.edu

Cerebellum (2012) 11:575–586DOI 10.1007/s12311-011-0319-5

While much is known about the effects of Met-Hg, onthe other hand, little is known about the developmentalneurotoxicity of Et-Hg, a metabolite of thimerosal (TM),used as a preservative in DPT vaccines in the USA until1999. TM is still found in flu vaccines administered topregnant women and infants in the USA and in DPTvaccines in developing countries. While the World HealthOrganization based its decision on the safety of TM invaccines on epidemiological studies, it has indicated thatthe data were derived from well-nourished, full-terminfants, and these data cannot necessarily be extrapolatedto preterm or malnourished infants [14]. As with otherenvironmental toxins, both health status and geneticallydetermined sensitivity to mercury may be crucial factors indetermining the overall outcome of exposure during thedevelopmental period.

The present study was undertaken to address thehypothesis that TM exposure during the perinatal periodimpairs cerebellar development by the mechanism involv-ing oxidative stress. TM was administered during G10–G15corresponding to the period of cerebellar Purkinje cell birthand simulating flu (and other TM containing) vaccinesgiven to pregnant women during the beginning of thesecond trimester of pregnancy [15] and P5–P10corresponding to the period of granule cells proliferationand a critical period of brain development; TM administra-tion during that time simulates vaccination during thesecond and the third trimester of pregnancy [15, 16]. To testthis hypothesis, we examined the effect of TM on neuro-developmental milestones, auditory functions, and motorlearning. We also examined cerebellar levels of theoxidative stress marker 3-nitrotyrosine (3-NT) and the type2 deiodinase (D2), a selenoenzyme that converts the pro-hormone thyroxine (T4) to the active hormone, 3′,3,5-triiodothyronine (T3) and is responsible for most of the T3supply within the brain [17]. These effects were examinedin male and female neonates to test for the sex-dependentnature of these effects and in two strains of rats withdifferent thresholds to oxidative stress, spontaneouslyhypertensive rats (SHR) and Sprague–Dawley (SD), to testfor genetically dependent sensitivity to Hg. We report herethat Hg exposure in the form of TM results in a variety ofneurodevelopmental deficits, altered cerebellar oxidativestress, and deiodinase activity, which are manifested in astrain- and sex-dependent manner.

Materials and Methods

Animals and Treatment

Timed-pregnant SHR or SD rat dams purchased fromCharles River Breeding Laboratories (Germantown, NY,

USA) on gestational day (G)7 (G1 defined as the first dayafter co-housing of males and females on which the femaleis found to have either a sperm plug or a sperm-positivevaginal smear) were individually housed under standardvivarium conditions (12:12 h light cycle, at 21–24°C).Standard laboratory chow and water were available adlibitum. Following a period of recovery from the stress ofshipment, selected SHR dams (n=3) and SD dams (n=6)received TM (Sigma-Aldrich, St. Louis, MO, USA) at adose of 200 μg/kg body weight (BW) via subcutaneousinjections from G10 through G15, and then again frompostnatal day (P)5 through P10; control SHR dams (n=3)and SD dams (n=3) received an equal volume of salinesolution injections.

The neonates in TM-exposed and control groups weretested for neurodevelopment milestones and auditory andmotor functions between birth and P20 and the cerebellartissue derived from these animals on P21 was analyzed foroxidative stress by measuring 3-NT levels, D2 activity, andgene expression. Maternal weight was monitored dailyfrom the onset of treatment. Neonates were counted, sexed,and weighed on P1/P2. Neonatal weight was monitoredfrom birth until euthanasia on P21. All procedures wereapproved by the Institutional Animal Care and UseCommittee at Harvard Medical School.

Neurodevelopmental Milestones

Neonatal neurodevelopmental milestones were assessedbetween birth and P21 (weaning) separately in male andfemale offspring derived from TM-exposed and controldams. Assessments included testing their righting re-sponse (rollover time), auditory (startle) response, andeye opening. Righting response was measured on P3–P5as the time required for a rat pup to right itself when placed ina supine position. Onset of the startle response—a sign ofacquiring hearing ability—was measured on P12–P14 interms of head movement response to the sound of pen tappedagainst a glass surface. Eye opening was recorded betweenP12 and P14.

Motor Functions

Motor functions were measured using a rotarod. One/twomales and one/two females per litter were chosen at randomfor the rotarod training test and these selected pups werefollowed over the period of 9 days, starting on P12 on arotarod with an accelerating speed setting through P20according to the procedure described earlier [18, 19]. Usingthis paradigm, motor learning was measured by increasingthe speed of rotation and evaluating the same neonates overtime; the pups tested on a rotarod from P12 through P20represented “trained pups”. The “trained group” included 7

576 Cerebellum (2012) 11:575–586

control SHR pups from three separate litters (3 males, 4females), 34 control SD pups from three litters (18 males,16 females), 6 TM-exposed SHR pups from three litters (3males, 3 females), and 77 TM-exposed SD pups from sixlitters (41 males, 36 females). Each neonate was subjectedto one trial on a rotarod rotating at incremental speeds inthe range of 2–20 rpm during 5-min intervals. The length oftime the animal remained on the rotarod and the rotationalspeed were recorded. If all animals were able to remain onthe rotarod during a 5-min interval, the speed of rotationwas increased. The cerebellar tissue derived from thesepups was fixed and not used for subsequent analysis, as thetraining itself may alter parameters measured in this studysuch as brain levels of 3-NT.

Remaining pups from each exposure group were kept asrotarod-naïve until P20 and were then tested on a rotarodset at a maximum speed of 20 rev/5 min. The pups testedfor the first time represented “untrained pups”. The“untrained” group included 6 control SHR pups from threelitters (3 males, 3 females), 34 control SD pups from threelitters (12 males, 10 females), 6 TM-exposed SHR pupsfrom three litters (3 males, 3 females) and 54 TM-exposedSD pups from six litters (30 males, 24 females).

Cerebellar Tissue

On P21 all pups were euthanized by decapitation. Thecerebella, from the pups tested on the rotarod only on P20,including the cerebellum, were rapidly dissected, frozen ondry ice, and stored at −80°C for further analysis.

Analysis of Cerebellar 3-NT Levels

The 3-NT levels were measured in the cerebellar homoge-nates prepared from frozen tissue according to the proce-dure previously described [20]. Briefly, the individualcerebella were homogenized in a phosphate buffer contain-ing detergents and protease inhibitors. The supernatantswere collected by centrifugation at 16,000×g for 30 min at4°C. 3-NT in the supernatants was measured using aliquots(equivalent to 25 mg tissue) derived from individual maleand female cerebellar samples (n=3–6 litters); 3-NT wasassayed with a commercially available 3-NT ELISA kit(Percipio Biosciences, Inc., Foster City, CA, USA). TheELISA plates were read at 450 nM. Data on 3-NT levels ofindividual cerebella were then pooled and the meansexpressed in picomoles per gram of tissue.

Analysis of Cerebellar D2 Activity

D2 activity was measured in the homogenates derived fromindividual male and female cerebellar samples by quantifying125I-release from a 125I labeled T4 tracer (5,700 mCi/mg,

Perkin Elmer Life Sciences, Boston, MA, USA) as describedpreviously [21]. For the assays, 50 μg of protein wereincubated for 4 h at 37°C with 1 nM T4 and 20 mM DTT.Background levels of deiodination were determined underidentical conditions using 100 nM unlabeled T4. Data on D2levels of individual cerebella were then pooled and expressedas means.

In Vitro Effect of Thimerosal Exposure in MouseEmbryonic Stem Cells

Cultures of Mouse Cortex E14 Neurospheres (StemCellTechnologies, Vancouver, BC Canada) were grown in theNeuroCult NSC Proliferation Kit (StemCell Technologies)according to the manufacturer’s directions. After twopassages, on day 4 in culture, TM was added to a finalconcentration of 1×10−7 M (170 nM); on day 6 in culture,both the control and TM neurospheres were harvested,dissociated using the NeuroCult Chemical Dissociation Kit(StemCell Technologies), and counted.

Analysis of Cerebellar TH-dependent Gene Expression

Cerebellar mRNA was isolated from randomly selected maleand female cerebella (n=4) per group using Trizol(Invitrogen, Carlsbad, CA, USA) following the manufac-turer’s instructions. Quantitative real-time PCR was used tomeasure gene expression levels and was performed asdescribed previously [19]; SuperScript VILO (Invitrogen,Carlsbad CA, USA) was used for cDNA synthesis followingthe manufacturer’s instructions. Sequences of primers usedare: rat Cyclophilin A 5′-AGCACTGGGGAGAAAGGATT-3′ and 5′-AGCCACTCAGTCTTGGCAGT-3′; rat Cirpb 5′-T CAGCT TCGACACCAATGAG - 3 ′ a n d 5 ′ -GTATCCTCGGGACCGGTTAT-3 ′ ; ra t Odf4 5 ′ -T T T T C C T C A C C C T C C T G T T G a n d 5 ′ -TGCAAGTAGCGTTGATGGAG-3′. PCR gene expressiondata was analyzed, corrected for cyclophilin expression, andthen pooled means were reported.

Statistical Analysis

The data presented here is derived from three to six littersper treatment group; both the number of litters and thenumber of male and female pups per groups is presented inthe context of each analysis. When applicable, a two-wayANOVAwas run with main effect of treatment and sex, andthe interaction between treatment and sex. If a statisticallysignificant interaction was found, test of treatment withineach sex was carried out. Because of the small sample size,power was too low for normality test. These comparisonswere thus repeated using Wilcoxon rank sum test and theresults were nearly identical. All values are reported as a

Cerebellum (2012) 11:575–586 577

mean ± the standard error of the mean. For all statisticaltests, p<0.05 was considered significant.

Results

Effect of TM Exposure During Pregnancy on Gainin Maternal Body Weight

For perinatal exposure, pregnant and lactating damsreceived 12 subcutaneous injections (Fig. 1) for a cumula-tive dose of 120 μg/dam. When given to a human infant ofan approximate weight of 5 kg as a 0.5 ml vaccine threedifferent times, the cumulative dose is 15 μg. While thedosage of TM per unit of body mass in the present studywas about 10 times higher than that used in humans, thelethal dose of Met-Hg at which fetal reabsorption could bedetected in rats was 8 mg/kg [22, 23]. There were no overtsigns of TM toxicity in SHR dams exposed duringpregnancy (G10–G15). The relative gain in maternal bodymass did not differ significantly between the TM exposedand control dams of either rat strain (data not shown).

Perinatal TM Exposure and Neonatal Weight Gain

SHR pups were much smaller than SD pups at birth(Table 1), and in general we observed a much greatermortality in TM-exposed SHR neonates (24%) than in TM-exposed SD neonates (4.3%); in SHR neonates attritionoccurred during the first 3 days after birth, while in SDpups it was mostly accounted by still births or deaths withinfirst few hours after birth As shown in Table 1, on P2, themass of TM-exposed SHR neonates was unexplainablysignificantly increased in both male (16.4% increase) and infemale pups (11.9% increase; p<0.01). The growth rate ofSHR pups was significantly suppressed by TM exposureboth in male and female pups and by P20 the weight of

TM-exposed neonates was slightly lower in both male andfemale pups.

In SD neonates, neonatal weight on P2 was not affected byTM exposure in either males or female neonates. Furthermore,the growth rate of TM-exposed SD pups was not affected andthere was no difference in weight between the TM exposedand control male or female pups at P21.

TM Exposure Does Not Affect Cerebellar Mass

While SHR pups were much smaller both at birth and on P21(Table 1), cerebellar mass of SHR neonates at P21 wascomparable to SD cerebellar mass. It appeared to be increasedin TM-exposed neonates, but the increase was not statisticallysignificant. Cerebellar mass of perinatally TM-exposed SDneonates assessed on P21 was not different (Table 1).

Effect of TM Exposure on Rollover Time

Unexplainably, the rollover time on P4 was shorter in SHRTM-exposed male pups (59%; p<0.05); females alsoshowed a tendency to decrease (13%) but this effect wasnot significant (Table 2). Notably, this also correlates with ahigher cerebellar and body mass at birth (Table 1) in TM-exposed SHR pups. The rollover time was not affected bythe TM exposure in SD male or female pups.

The Effect of TM Exposure on the Onset of the AuditoryResponse

The maturation of the auditory functions was tested bystartle response (Table 2). In the SHR pups exposed to TM,the startle response measured on P14 showed a smallsuppression in males 12.2% but not in female pups.However, in the SD strain, the percent of male pupsshowing a startle response was significantly reduced by27.8% and in female pups by 19.2%.

SNOITCEJNIMTSNOITCEJNIMT

Rat G10 G15 Birth P4 P5 P10 P12 P14 P20 P21

OE,TSOR

ROTAROD TR

ROTAROD UNTRPURKIJE CELL BIRTH GRANULE CELL PROLIFERATION

CRITICAL BRAIN DEVELOPMENT PERIOD

Human 1st/2nd trimester 3rd trimester

Fig. 1 Schematic representation of thimerosal (TM) exposure andneurobehavioral testing. Timed pregnant spontaneously hypertensive rats(SHR) or Sprague–Dawley (SD) received TM at a dose of 200 μg/kgbody weight via subcutaneous injections during G10–G15 and P5–P10;control SHR and SD dams received an equal volume of saline. Rollovertime (RO) was measured on P3–P5, startle response (ST), and eye

opening were observed on P12–P14. Male and female pups were testedon a rotarod (RT) either on P20 to measure spontaneous motor function inuntrained group (ROTAROD UNTR) or between P12 and P20 to assessthe learned motor function in trained pups (ROTAROD TR). Animalswere euthanized on P21 and cerebella dissected out for biochemicalanalysis.

578 Cerebellum (2012) 11:575–586

TM Exposure Does Not Affect Eye Opening

Eye opening in TM-exposed SHR males showed atendency to be delayed, while no effect was observed inTM-exposed SHR females. No effect of TM exposure oneye opening was observed in the SD strain (Table 2).

TM Exposure Impairs Motor Learning

Effect of TM Exposure on Spontaneous Motor Functionon P20

To assess the spontaneous motor behavior, a subset ofrotarod “naïve” rats was tested on P20 on a rotarod set to amaximum setting of 20 rpm/5 min; this subset is referred to

as “untrained pups” (Fig. 2). TM exposure did not affectfalling latency in untrained SHR male (Fig. 2a) or femalepups (Fig. 2b). Similarly, perinatal TM exposure did notaffect falling latency in untrained SD male (Fig. 2c) or SDfemale pups (Fig. 2d).

Effect of TM Exposure on Motor Learning

To assess the effect of TM on motor learning, selected maleand female neonates from each litter were tested on arotarod daily commencing on P12 and continuing until P20;this subset is referred to as “trained pups” (Fig. 2).

Overall, trained SHR pups had poorer performance on arotarod, and their falling latency was shorter than thatobserved in SD pups (Fig. 2). TM-exposed SHR male pups

Table 1 Effect of perinatal TM exposure on neonatal growth and cerebellar weight

Strain Sex Treatmentgroup (n)

Neonatalmass on P2 (g)

Neonatalmass on P21 (g)

Relative growth(P21–P2)/P2 (%)

Cerebellar masson P21 (mg)

SHR Male C (3) 5.35±0.05 27.60±1.20 4.16±0.27 161.13±21.38

TM (3) 6.23±0.19 ** 26.70±1.30 3.28±0.14* 173.64±3.39

Female C (3) 5.30±0.10 27.00±0.70 4.10±0.23 153.75±9.75

TM (3) 5.93±0.2** 26.60±2.4 3.48±0.19* 173.06±1.47

SD Male C (3) 10.07±0.38 49.83±1.18 3.96±0.0.08 171.11±6.57

TM (7) 10.04±0.24 51.73±1.87 4.15±0.15 173.46±2.96

Female C (3) 9.57±0.26 48.37±0.67 4.06±0.15 165.83±10.25

TM (7) 9.46±0.28 48.50±1.92 4.22±0.08 161.67±3.53

The neonatal mass on P2 was increased in TM-exposed SHR males and females, but the growth rate was actually decreased. Neonatal mass wasnot affected by TM exposure in SD neonates. Cerebellar mass was not significantly affected by TM exposure in either strain. In the analysis ofneonatal and P21 mass and the cerebellar weight, litter was used as a unit of observations; the numbers presented here are averages across litters

C controls, TM thimerosal-exposed, l number of litters, n number of neonates

*p<0.05; **p<0.01

Table 2 Effect of perinatal TM exposure on neurodevelopmental milestones

Strain Sex Treatment group (n) Rollover time (s) Startle response (% responders) Eye opening (% of pups)

SHR Male C (3) 14.2±7.0 73.3±6.7 100.0±0

TM (3) 6.4±1.5 * 64.3±18.0 83.3±16.7

Female C (3) 14.6±2.9 87.5±12.5 100.0±0

TM (n=3) 12.7±2.9 91.7±8.3 100.0±0

Male C (n=3) 7.1±2.8 74.1±25.9 66.7±33.3

SD TM (7) 6.9±2.5 53.5±7.8 * 78.2±10.6

Female C (3) 11.4±2.0 90.5±9.5 85.7±14.3

TM (7) 13.5±2.2 73.1±9.1 * 83.9±8.1

The rollover time on P4 was shorter in SHR TM exposed male, but not female neonates; it was not affected in SD neonates. TM exposure did notaffect the number of SHR neonates of either sex responding to auditory stimulus; but it reduced the number of responding SD males and females.Eye opening was delayed in SHR rats by ~3 days. On P17, TM exposed SHR male, but not female neonates, showed a tendency towards delayedeye opening on. No effect of TM exposure on eye opening was observed in SD pups on P14. In the analysis of rollover, startle response and eyeopening, litter was used as a unit of observations; the numbers presented here are averages across litters

C controls, TM thimerosal-exposed, n number of litters

*p<0.05

Cerebellum (2012) 11:575–586 579

falling latency was 77.4% of that observed in control malepups (p<0.05; Fig. 2a), while in TM-exposed female pupsthere was no difference from the control group (Fig. 2b).

Perinatal TM exposure induced motor impairment in SDpups of both sexes. The mean falling latency in TM-exposed male pups was 70.2% of that observed in controlmale pups (p<0.05; Fig. 2c), and in TM-exposed femalepups it was 45.0% of that observed in control female pups(p<0.05; Fig. 2d).

TM Exposure Results in Increased Cerebellar Levelsof Oxidative Stress Marker, 3-NT

The levels of the oxidative stress marker 3-NT werequantified in cerebellar tissue obtained from TM-exposedmale and female SHR and SD rat pups on P21 (Fig. 3). Thelevels of 3NT were increased in both male (6.2%; Fig. 3a)and in female SHR neonates (35.1%; p<0.05; Fig. 3b)exposed to TM. Exposure to TM in SD rats resulted in14.0% increase in 3-NT in male neonates (p<0.05; Fig. 3c)but there was no significant change in the female neonates(Fig. 3d).

TM Exposure Results in Decreased Cerebellar D2 Activity

The levels of D2 activity were quantified in cerebellartissue from control and TM-exposed P21 male and femaleSHR and SD rat neonates (Fig. 4). Perinatal exposure toTM resulted in a very significant 60.9% decrease incerebellar D2 activity in male (p<0.01; Fig. 4a), but notin female SHR neonates (Fig. 4b). TM exposure alsotended to decrease D2 activity in male and female SDneonates (Fig. 4c, d).

TM Exposure Disrupts TH-dependent Gene Expression

The decrease in D2 activity observed in SHR male neonatessuggests that there may be less local T4 to T3 conversion inthe cerebellum of these animals resulting in a decreased T3content within this tissue (Fig. 4a). To assess if this affectsdownstream T3-regulated gene expression, we determinedthe mRNA levels of two genes negatively regulated by T3that have been previously shown to be upregulated in D2KOmice [24]. Expression of Cirbp remained unchanged in TM-exposed SHR males and females (Fig. 5a, b). Notably, Odf4

SD MALES

a.

c.

SHR MALES b.

d.

SHR FEMALES

SD FEMALES

*

*

*

0

50

100

150

200

250

P12 P20-TR P20-UNTR

RO

TAR

OD

TIM

E

(sec

)

CTM

0

50

100

150

200

250

P12 P20-TR P2-UNTR

RO

TAR

OD

TIM

E

(sec

) CTM

*

0

100

200

300

400

P12 P20-TR P20-UNTR

RO

TAR

OD

TIM

E (

sec) C

TM

*

0

100

200

300

400

P12 P20-TR P20-UNTR

RO

TAR

OD

TIM

E (

sec)

CTM

*

Fig. 2 TM exposure impairs motor learning. Motor performance ispresented as falling latency on the rotarod apparatus (seconds). To assessthe spontaneous motor behavior a rotarod naïve “untrained pups” (P20-UNTR) were tested on P20 on a rotarod set to a maximum setting of20 rpm/5 min. For this analysis, litter was used as a unit of observation;data is presented as means±SEM. The untrained pups are represented by:a C n=3, TM n=3; b C n=3, TM n=3; c C n=3, TM n=6; and d C n=3, TM n=6. Perinatal TM exposure did not affect falling latency in SHRmales (Fig. 1a) or female pups (Fig. 1b). Similarly, perinatal TMexposure did not affect falling latency in SD male (Fig. 1c) or SDfemale pups (Fig. 1d). To assess the effect of TM on motor learning,

selected male and female neonates from each litter were tested on arotarod daily commencing on P12 neonates and continued until P20(P20-TR). The trained pups are represented by: a C three litters, threemales; TM three litters, three males; b C three litters, four females; TMthree litters, three females; c C three litters, 18 males; TM six litters, 41males; d C three litters, 16 females; TM six litters, 36 females. Themean falling latency in SHR TM-exposed male pups decreased by22.6% (Fig. 1a; *p<0.05), and in SHR TM-exposed female pups itdecreased by 6.6% (Fig. 1b). The mean falling latency in TM-exposedSD male pups was decreased by 29.8% (Fig. 1c; *p<0.05), and in TM-exposed SD female pups it was decreased by 55% (Fig. 1d; *p<0.05)

580 Cerebellum (2012) 11:575–586

expression was increased by 47.7 % in TM-exposed males(p<0.05; Fig. 5a) consistent with the decreased D2 activityfound in this group that likely resulted in lower T3 levelswithin the cerebellum of these animals.

TM Induces Cell Apoptosis in Embryonic Stem CellsIn Vitro

The effect of in vitro TM exposure on mouse embryonic stemcells is shown in Fig. 6. Addition of TM to a finalconcentration of 1.7×10−7 M, similar to one used in primary

rat cerebellar [25] and human blood mononuclear cells incultures [26], resulted in a 65.5 % reduction (p<0.05; Fig. 6)on a number of viable cells within 48 h. The number ofviable cells in TM cultures fell below cell number prior toTM addition, indicating the observed reduction in cellnumber was due to decreased survival of neuronal cells.

Limitations of the Study

The experimental design of the present study incorporatesattempts to model both the perinatal Et-Hg/TM exposure

0

50

100

150

200

250

300

350

3-N

T (

pmol

/g)

b. SHR FEMALESa. SHR MALES

d. SD FEMALESc. SD MALES

0

50

100

150

200

250

300

350

3-N

T (

pmol

/g)

C TM

0

50

100

150

200

250

3NT

(pm

ol/g

)

*

C TM

C TM

0

50

100

150

200

250

3NT

(pm

ol/g

)

C TM

*

Fig. 3 TM exposure results inincreased cerebellar 3-NT levels.Cerebellar 3-NT levels werepresented as picomoles per gramtissue. a C Three litters, sixmales; TM three litters, 14males. b C Three litters, sevenfemales; TM three litters, sevenfemales. c C three litters, sevenmales; TM six litters, 16 males.d C Three litters, six females;TM three litters, 15 females.Data are presented as mean±SEM. The levels of 3NT wereincreased in both SHR male (a,6.2%) and SHR female neonates(b, *p<0.05) exposed to TM.Exposure to TM in SD ratsresulted in an increase in 3-NTin male neonates (c, *p<0.05),but there was no significantchange in the female neonates(Fig. 2d)

0

0.2

0.4

0.6

0.8

D2

act

ivity

(fm

ol/m

in/m

g)

C TM

SD MALES

SHR MALESa.

c.

b.

d.

SHR FEMALES

SD FEMALES

0

0.2

0.4

0.6

0.8

D2

activ

ity (

fmo/

min

l/mg)

C TM

0

1

2

3

4

D2

activ

ity (

fmol

/min

/mg)

C TM0

1

2

3

4

D2

activ

ity (

fmo/

min

l/mg)

C TM

**

Fig. 4 TM exposure results indecreased cerebellar D2 activity.Cerebellar D2 activity is pre-sented in femtomole per minuteper milligram of tissue. a Cthree litters, five males; TMthree litters, 14 males. b C Threelitters, eight females; TM threelitters, seven females. c C threelitters, four males; TM six litters,eight males. d C Three litters,four females; TM six litters, fivefemales. Data are expressed asmean±SEM. Perinatal exposureto TM resulted in decreasedcerebellar levels of D2 in SHRmale (Fig. 3a; **p<0.01) butnot in SHR female neonates(Fig. 3b). On the other hand, D2was decreased both in TM-exposed SD males (Fig. 3c) andSD female neonates (Fig. 3d)

Cerebellum (2012) 11:575–586 581

and exposure to TM-containing vaccines administeredduring pregnancy on the developing brain. Somehowunorthodox schedule of repeated TM injections attemptsto model a scenario of maternal exposure to flu and severalother TM-containing vaccines such as hepatitis B, pneu-mococcal, meningococcal, and rabies vaccines recommen-ded to high-risk mothers. While the present paradigm mayexaggerate human exposure, it forms basis for future in-depth studies.

By including two strains of rats, SD and SHR, withdifferential sensitivity to environmental triggers of oxida-tive stress [33–35], we have addressed the individualgenetic susceptibility to environmental toxicants, whichmay ultimately determine the consequences of the exposureto TM. The limited number of SHR dams (due to thefinancial constrains) as compared to SD dams, couldpotentially influence the reported outcomes and limit theinterpretation of the study.

Another factor that could influence the outcomes of ourstudy is the decision not to cull litters. Litter culling is a

controversial practice which offers both advantages anddisadvantages. Culling of the rodent litters has beenrecommended by some [66] to reduce variability in thegrowth and development of pups. On the other hand, others[67] are highly critical of culling claiming lack of evidencefor clear advantage of the process that eliminates 30–45%of pups and risks missing the effects of neurotoxicity, suchas pups mortality and malformations. Indeed, in the case ofSHR pups the attrition could be observed during the first3 days.

Discussion

Epidemiological surveys have shown that the industrialrelease of mercury into Minamata Bay in Japan andconsumption of Met-Hg contaminated fish resulted in adistinct pathology recognized as Minamata diseaseaffecting neonates and children with the brain being aprimary target organ [27]. Met-Hg exposure in expectingmothers due to fish consumption is associated withincreased mercury accumulation in the infant brainsaccompanied by behavioral abnormalities, which includedeficits in motor, attention, and verbal performance thatare more pronounced in males [28], while the postnatalMet-Hg exposure in humans appears to have no recognizableeffects [29].

Hg is transferred through the placenta, with theconcentration in fetal blood cells being 30% higher thanin maternal blood cells [30]. Studies in mice have shownthat gestational exposure (G8–G18) through a diet contam-inated with Met-Hg (100 μg/kg) resulted in a deficit inmotor behavior and coordination [31]. Hg also enters themilk [32], is taken up by the suckling pups [6], andaccumulates in their brains [4, 33]. In rats, postnatal

a. SHR MALES b. SHR FEMALES

0

0.5

1

1.5

2

2.5

3

Cirbp Odf4

RE

LAT

IVE

EX

PR

ES

SIO

N

C TM C TM

0

0.5

1

1.5

2

2.5

3

Cirbp Odf4

RE

LAT

IVE

EX

PR

ES

SIO

N*

Fig. 5 TM exposure disrupts TH-dependent gene expression. Cerebellargene expression was measured by quantitative RT-PCR and wasnormalized to cyclophilin A expression. a C Cirbp, three litters, fourmales; C Odf4, three litters, four males; TM Cirbp three litters four males;TM Odf4 three litters, four males. b C Cirbp three litters, four females; C

Odf4 three litters, four females; TM Cirbp three litters four females; TMOdf4 three litters, four females. Data are presented as a relative geneexpression (mean±SEM). An increase in Cirbp expression was observedin TM-exposed SHR females (b) and an increase in Odf4 expression wasobserved in TM-exposed SHR male cerebella (a, *p<0.05)

0

1

2

3

4

5

6

7

TO

TAL

CE

LL S

X10

e7

C TM

*

Fig. 6 TM induces cell apoptosis in embryonic stem cells in vitro.Mouse embryonic stem cells are presented as total number of cells perwell; data are expressed as mean±SEM. Exposure to TM (1.7×10−7 M) resulted in a significant reduction (*p<0.05) in the number ofviable cells within 48 h

582 Cerebellum (2012) 11:575–586

exposure (P1–P30) resulted in impairments in motorcoordination and learning [34]. It is important to point outthat our exposure paradigm overlaps with the critical periodof cerebellar development in rats (P5–P10; Fig. 1).

While many toxicological studies have documented thedevelopmental neurotoxicity of Met-Hg, few studies haveexplored the impact of Et-Hg/TM on the developing centralnervous system (CNS). The immediate objective of thepresent study was to assess the neurodevelopmentalconsequences of perinatal Et-Hg/TM exposure, while theultimate goal has been to gain insight into the potentialimpact of TM-containing vaccines administered duringpregnancy on the developing human brain. Thus, theexperimental design of the present study incorporatesattempts to model several aspects relevant to both the shortand the long term objectives. The paradigm of repeated TMinjections attempts to model administration of flu and otherthimerosal-containing vaccines such as hepatitis B, pneu-mococcal, meningococcal, and rabies vaccines recommen-ded to the high-risk mothers. Thus, while the multiple TMadministrations may exaggerate human exposure, they formthe basis for future more in-depth studies relevant to asingle flu vaccine during pregnancy. This study alsoaddresses the issue of the individual genetic susceptibilityto environmental toxicants, by including two strains of rats,SD and SHR, with differential sensitivity to environmentaltriggers of oxidative stress [35–37]. Furthermore, the studyaddresses sex-dependent nature of environmental impactson the developing CNS by including a separate analysis ofTM effects in male and in female neonates. This study alsoattempts to link the environmental exposure to TM duringG10–G15 and P5–P10 not only to specific periods ofhuman pregnancy, i.e., the beginning of the secondtrimester and the third trimester, respectively, but also tocritical developmental events, i.e., Purkinje cell birth andgranule cell migration.

Our results indicate that perinatal TM exposure delaysauditory maturation [38] and impairs motor learning [34],consistent with previous results observed in rodentsexposed to Met-Hg. Our data show that the percentage ofboth male and female SD neonates displaying a startleresponse at 14 days with TM exposure was significantlydecreased, indicating a deficit in auditory development. Wefurther found that spontaneous motor coordination inneonates of either sex or strain measured on P20 was notaffected by perinatal TM exposure. However, notably, TMexposure does appear to affect learned motor function. Inthe SHR pups, which appear to be less coordinated overall, learned motor function was impaired in male pups.On the other hand, the learned motor coordination wasimpaired in SD rats of both sexes, with a larger effect infemale neonates. While comparative studies on theeffects of TM are lacking, our results are in agreement

with the previously published studies of the Met-Hgeffect on learning [22, 31, 34].

Impaired cerebellar development due to Hg exposure hasbeen suggested by several studies. Combined gestationaland neonatal Met-Hg exposure in rats results in increasedheight of Purkinje cells [39], while postnatal exposure isassociated with neuronal degeneration and astrocytosis inthe cortex, striatum, and the cerebellum [34]. In neonatalcerebellar neurons in culture, both Met-Hg and TM increaseintracellular calcium concentration and exert a cytotoxiceffect on granule cells [40]. Our in vitro results involvingembryonic neuronal stem cells indicate that the immatureneurons are very sensitive to TM, with exposure resultingin decreased cell proliferation/survival (Fig. 6).

Like many other environmental toxicants, Hg accumu-lated in the developing brain following either in vivo Met-Hg exposure [41] or in vitro exposure to Met-Hg or Et-Hg[41–43], induces oxidative stress that leads to a cascade ofother changes including decreased neurogenesis, increasedneuronal apoptosis and impaired synaptic plasticity in theneonatal brain. Further, gestational exposure to Met-Hg inmice results in increased lipid peroxidation via interferencebrain GSH levels [13], while gestational exposure (G12–G14) in rats to Met-Hg (5 mg/kg) induces oxidative stressand reduces the antioxidant enzyme superoxide dismutasein the hippocampus [44]. However, a direct quantification ofoxidative stress is impractical because free radicals includingreactive oxygen species are short-lived. An alternativeapproach is to monitor stable end-products of oxidative stressdamage. 3-NT is a well-accepted marker of oxidative stressfound in over 50 different pathologies [45] includingAlzheimer’s and Parkinson’s diseases [46, 47] and autism[20]. Results of our present study indicate that perinatal TMexposure also increases cerebellar oxidative stress, as assessedby 3-NTmeasurement. The effect is more pronounced in SHRpups, with the increase observed in both male and femaleneonates, but significant only in female pups. In SD neonates,increased 3-NT levels were only observed in males.

We also found that TM exposure significantly decreasedD2 activity in the cerebellar tissue of SHR male neonates,and also showed a tendency to decreased D2 activity inmale and female SD neonates. The D2 enzyme catalyzesthe activation of the pro-hormone T4 to the biologicallyactive T3, and plays a key role in the local control ofthyroid hormone levels within a tissue [48]. Further, it hasbeen estimated that greater than 80% of the T3 foundwithin the brain comes from local T4 to T3 conversion byD2 [17], and mice with a global targeted disruption of theDio2 gene [D2KO mice] have ~50% less T3 content intheir cerebral cortex, cerebellum, and hypothalamus [49].Thus, a decrease in D2 activity within the cerebella of TM-exposed SHR male neonates potentially could result inlocal cerebellar “hypothyroidism”. Such local decrease in

Cerebellum (2012) 11:575–586 583

TH adds yet another possible mechanism of action to themercury compounds already qualified as endocrine dis-ruptors [2, 3, 9, 10].

Furthermore, D2KO mice have been found to havesignificantly increased expression of numerous genes thatare negatively regulated by T3 [24]. Consistent with this,we find that expression of the one of the T3-responsivemarker genes found to be altered in D2KO mice, Odf4, isalso increased in TM-treated SHR male neonates, repre-senting the first evidence of altered TH-dependent geneexpression following TM exposure. This suggests that Et-Hg may affect local brain T3 levels that are critical fornormal neurodevelopmental processes, in turn impactingcerebellar development and impairing cerebellum-dependent auditory and motor functions. Indeed, our dataon auditory and motor functions support the hypothesis ofT3 deficiency in TM-exposed rat neonates. Interestingly,impairment of auditory and motor functions is alsoobserved in propylthiouracil-induced hypothyroid condi-tions in rats [50], and D2KO mice have mildly reducedmotor function and learning deficits in some, but not all,tests [49]. Auditory impairment can be observed in D2knockout mice, however this is due to defects in cochleardevelopment [51].

While the mechanism by which D2 activity is decreasedin TM-treated neonates is unclear, Met-Hg also interactswith selenium [52] and can inhibit function of selenopro-teins such as the deiodinases [53]. We have also shown thatTM exposure increases levels of oxidative stress, which hasbeen found previously to decrease expression of the Dio2gene [54]. Lastly, exposure of neuronal cells to Met-Hg[55] or neuroblastoma cells to TM [56] results in adepletion of GSH which is both an antioxidant and acofactor of deiodinases [57–59], thus cerebellar D2 activitymight be impaired due to a lack of reducing co-factor.Additionally, T3 regulates GSH levels in the developingbrain and treatment of astrocyte cultures with T3 results inincreased GSH levels and improved antioxidative defense,suggesting that TH plays a positive role in maintainingGSH homeostasis and protecting the brain from oxidativestress [60]. Thus, is it is also possible that a decrease in D2activity could further amplify the effects of oxidative stress.

Our behavioral and biochemical data indicate that theeffects of perinatal TM exposures are sex-dependent. Thesefindings are at least in part in agreement with earlierobservations both in humans [28] and in animals [61]showing that the developing males appear to be moresensitive to Hg exposure. The present study indicates thatlearned motor coordination is impaired more significantlyboth in SD and SHR males, but is unaffected in SHRfemales. Our results also indicate that perinatal TMexposure induces an increase in cerebellar oxidative stressin SD males but not females; however in SHR rats, the

increase is greater in females. Perinatal exposure to TMresulted in decreased cerebellar D2 activity in male, but notin female SHR neonates, and this decrease was correlated witha disruption of T3-dependent gene expression in SHR males.

In addition to the sex-dependent effect of TM exposureon the developing CNS, the effect is also distinct indifferent rat strains. Other studies reported strain differencesin Hg effect on sensitivity to pain [62] and renal HGtoxicity [63]. This difference may be attributed to geneti-cally dependent susceptibility to environmental toxicantsincluding Hg, although the mechanisms involved are poorlyunderstood [64]. Our results show higher levels of the brainoxidative stress marker 3-NT in SHR than in SD rats, whichis consistent with previous reports [35–37], that may bedue, at least in part, to a different degree of activation ofinflammatory processes between the two strains [65].

Conclusions

Our data indicate that maternal TM exposure results in adelayed auditory maturation and impaired motor learning inrat pups. Factors that may contribute to these abnormalitiesinclude increased cerebellar oxidative stress and decreasedD2 activity resulting local intracerebellar T3 deficiency andaltered TH-dependent gene expression. Indeed, provided hereis the first evidence of altered TH-dependent gene expressionfollowing TM exposure. Our data thus demonstrate a negativeneurodevelopmental impact of perinatal TM exposure, whichappears to be both strain- and sex-dependent. Although,additional studies are needed, data derived from TM exposurein rats may provide clues relevant to understanding neuro-developmental consequences of TM exposure in humans.

Acknowledgments We would like to thank the Mercury as a GlobalHazard SGIG for the grant awarded by the College of William and Maryto Z.L. Sulkowski, the Autism Research Institute and SafeMinds forgrants awarded to Dr. Sajdel-Sulkowska, and the NIDDK-DK76117grant awarded to A.M. Zavacki. We would also like to thank Puja Parekhof the College of William for participating in the initial experiments andMing Xu, Dept. Integrative Physiology, Gunma University GraduateSchool of Medicine, Maebashi Gunma, Japan, for initial tissue analysis.We also acknowledge the following Sponsored Research Staff membersat Brigham and Women’s Hospital: Amrutha E. Mathew, Pooja Mathewand Ashesh Shresta for RNA preparation, and Dr. Alaptagin Khan forPCR primer validation.

References

1. Bokara KK, Brown E, McCormick R, Yallapragada PR, RajannaS, Bettaiya R. Lead-induced increase in antioxidant enzymes andlipid peroxidation products in developing rat brain. Biometals.2008;21:9–16.

2. Windham GC, Zhang L, Gunier R, Croen LA, Grether JK. Autismspectrum disorders in relation to distribution of hazardous air

584 Cerebellum (2012) 11:575–586

pollutants in the San Francisco Bay Area. Environ HealthPerspect. 2006;114:1438–44.

3. Palmer RF, Blanchard S, Wood R. Proximity to point sources ofenvironmental mercury release as a predictor of autism preva-lence. Health Place. 2009;15:18–24.

4. Orct T, Blanusa M, Lazarus M, Varnai VM, Kostial K.Comparison of organic and inorganic mercury distribution insuckling rat. J Appl Toxicol. 2006;26:536–9.

5. Nordenhäll K, Dock L, Vahter M. Transplacental and lactationalexposure to mercury in hamster pups after maternal administration ofmethyl mercury in late gestation. Pharmacol Toxicol. 1995;77:130–5.

6. Oskarsson A, Palminger Hallén I, Sundberg J. Exposure to toxicelements via breast milk. Analyst. 1995;120:765–70.

7. Glaser V, Nazari EM, Muller YM, Feksa L, Wannmacher CM,Rocha JB, et al. Effects of inorganic selenium administration inmethylmercury-induced neurotoxicity in mouse cerebral cortex.Int J Dev Neurosci. 2010;28:631–7.

8. Yin ZZ, Lee E, Ni M, Jiang H, Milatovic D, Rongzhu L, et al.Methylmercury-induced alterations in astrocyte function areattenuated by ebselen. Neurotoxicology. 2011;32(3):291–9.

9. Heath JA, Frederick PC. Relationship among mercury concen-trations, hormones, and nesting effort of white Ibises (Eudocimusalbus) in the Florida Everglades. Auk. 2005;122:255–67.

10. Tan SW,Meiller JC, Mahaffey KR. The endocrine effects of mercuryin humans and wildlife. Crit Rev Toxicol. 2009;39:228–69.

11. Chang JY, Tsai PF. Prevention of methylmercury-induced mito-chondrial depolarization, glutathione depletion and cell death by15-deoxy-delta-12,14-prostglandin J(2). Neurotoxicology.2008;29:1054–61.

12. Barcelos GR, Grotto D, Serpeloni JM, Angeli JP, Rocha BA, et al.Protective properties of quercetin against DNA damage andoxidative stress induced by methylmercury in rats. Arch Toxicol.2011;85(9):1151–7.

13. Stringari J, Nunes AK, Franco JL, Bohrer D, Garcia SC, Dafre AL, etal. Prenatal methylmercury exposure hampers glutathione antioxi-dant system ontogenesis and causes long-lasting oxidative stress inthe mouse brain. Toxicol Appl Pharmacol. 2008;227:147–54.

14. WHO. The Global Advisory Committee on Vaccine Safety,Statement on thiomersal. http://www.who.int/vaccine_safety/topics/thiomersal/statement_jul2006/en/index.html; 2006.

15. Maier SE, Cramer JA, West JR, Sohrabji F. Alcohol exposureduring the first two trimesters equivalent alters granule cellnumber and neurotropin expression in the developing rat olfactorybulb. J Neurobiol. 1999;41:414–23.

16. Bellinger FP, Bedi KS, Wilson P, Wilce PA. Ethanol exposureduring the third trimester equivalent results in long-lastingdecreased synaptic efficacy but not plasticity in the CA1 regionof the rat hippocampus. Synapse. 1999;31:51–8.

17. Silva JE, Leonard JL, Crantz FR, Larsen PR. Evidence for twotissue-specific pathways for in vivo thyroxine 5′-deiodination inthe rat. J Clin Invest. 1982;69:1176–84.

18. Nguon K, Baxter MG, Sajdel-Sulkowska EM. Perinatal exposureto polychlorinated biphenyls differentially affects cerebellardevelopment and motor functions in male and female rat neonates.Cerebellum. 2005;4:112–22.

19. Sajdel-Sulkowska EM, Nguon K, Sulkowski ZL, Rosen GD,Baxter MG. Purkinje cell loss accompanies motor impairment inrats developing at altered gravity. Neuroreport. 2005;16:2037–40.

20. Sajdel-Sulkowska EM, Lipinski B, WindomH, Audhya T, McGinnisW. Oxidative stress in autism: cerebellar 3-nitrotyrosine levels. Am JBiochem Biotechnol. 2008;4:73–84.

21. Zavacki AM, Ying H, Christoffolete MA, Aerts G, So E, HarneyJW, et al. Type 1 iodothyronine deiodinase is a sensitive marker ofperipheral thyroid status in the mouse. Endocrinology.2005;146:1568–75.

22. Eccles CU, Annau Z. Prenatal methyl mercury exposure: I. Alterationsin neonatal activity. Neurobehav Tocicol Teratol. 1982a;4:371–6.

23. Eccles CU, Annau Z. Prenatal methyl mercury exposure: II.Alterations in learning and psychotropic drug sensitivity in adultoffspring. Neurobehav Tocicol Teratol. 1982;4:377–82.

24. Morte B, Ceballos A, Diez D, Grijota-Martinez C, DumitrescuAM, Di Cosmo C, et al. Thyroid hormone-regulated mousecerebral cortex genes are differentially dependent on the source ofthe hormone: a study in monocarboxylate transporter-8- anddeiodinase-2-deficient mice. Endocrinology. 2010;151:2381–7.

25. Zieminska E, Toczylowska B, Stafiej A, Lazarewicz JW. Lowmolecular weight thiols reduce thimerosal neurotoxicity in vitro:modulation by proteins. Toxicology. 2010;276:154–63.

26. Gardner RM, Nyland JF, Silbergeld EK. Differential immunotoxiceffects of inorganic and organic mercury species in vitro. ToxicolLett. 2010;198:182–90.

27. Clarkson TW, Magos L, Myers GJ. The toxicology of mercury—current exposures and clinical manifestations. N Eng J Med.2003;349:1731–7.

28. Gao Y, Yan CH, Tian Y, Xie HF, Zhou X, Yu XD, et al. Prenatalexposure to mercury and neurobehavioral development of neo-nates in Zhoushan City, China. Environ Res. 2007;105:390–9.

29. Debes F, Budtz-Jørgensen E, Weihe P, White RF, Grandjean P.Impact of prenatal methylmercury exposure on neurobehavioralfunction at age 14 years. Neurotoxicol Teratol. 2006;28:363–75.

30. Kuhnert PM, Kuhnert BR, Erhard P. Comparison of mercurylevels in maternal blood, fetal cord blood, and placental tissue.Am J Obstet Gynecol. 1981;139:209–13.

31. Montgomery KS, Mackey J, Thuett K, Ginestra S, Bizon JL,Abbott LC. Chronic, low-dose prenatal exposure to methylmer-cury impairs motor and mnemonic function in adult C57/B6 mice.Behav Brain Res. 2008;191:55–61.

32. Morgan DL, Price HC, Fernando R, Chanda SM, O’Connor RW,Barone Jr SS, et al. Gestational mercury vapor exposure and dietcontribute to mercury accumulation in neonatal rats. EnvironHealth Perspect. 2006;114:735–9.

33. Zareba G, Cernichiari E, Hojo R, Nitt SM, Weiss B, Mumtaz MM,et al. Thimerosal distribution and metabolism in neonatal mice:comparison with methyl mercury. J Appl Toxicol. 2007;27:511–8.

34. Sakamoto M, Kakita A, de Oliveira RB, Sheng Pan H, Takahashi H.Dose-dependent effects of methylmercury administered during neona-tal brain spurt in rats. Brain Res Dev Brain Res. 2004;152:171–6.

35. Kodavanti UP, Schladweiler MC, Ledbetter AD, Ortuno RV,Suffia M, Evansky P, et al. The spontaneously hypertensive rat: anexperimental model of sulfur dioxide-induced airways disease.Toxicol Sci. 2006;94:193–205.

36. Wang X, Desai K, Juurlink BH, de Champlain J, Wu L. Gender-related differences in advanced glycation endproducts, oxidativestress markers and nitric oxide synthases in rats. Kidney Int.2006;69:281–7.

37. Saiki R, OkazakiM, Iwai S, Kumai T, Kobayashi S, Oguchi K. Effectsof pioglitazone on increases in visceral fat accumulation and oxidativestress in spontaneously hypertensive hyperlipidemic rats fed a high-fatdiet and sucrose solution. J Pharmacol Sci. 2007;105:157–67.

38. Beyrouty P, Stamler CJ, Liu JN, Loua KM, Kubow S, Chan HM.Effects of prenatal methylmercury exposure on brain monoamineoxidase activity and neurobehaviour of rats. Neurotocicol Teratol.2006;28:251–9.

39. Roegge CS, Morris JR, Villareal S, Wang VC, Powers BE,Klintsova AY, et al. Purkinje cell and cerebellar effects followingdevelopmental exposure to PCBs and/or MeHg. NeurotocicolTeratol. 2006;28:74–85.

40. Ueha-Ishibashi T, Oyama Y, Nakao H, Umebayashi C, NishizakiY, Tatsuishi T, et al. Effect of thimerosal, a preservative invaccines, on intracellular Ca2+ concentration of rat cerebellarneurons. Toxicology. 2004;195:77–84.

Cerebellum (2012) 11:575–586 585

41. Linares AF, Loikkanen J, Jorge MF, Soria RB, Novoa AV.Antioxidant and neuroprotective activity of the extract from theseaweed, Halimeda incrassata (Ellis) Lamouroux, against in vitroand in vivo toxicity induced by methyl-mercury. Vet HumToxicol. 2004;46:1–5.

42. Kaur P, Aschner M, Syversen T. Glutathione modulation influencesmethyl mercury induced neurotoxicity in primary cell cultures ofneurons and astrocytes. Neurotoxicology. 2006;27:492–500.

43. Rush T, Hjelmhaug J, Lobner D. Effects of chelators on mercury,iron, and lead neurotoxicity in cortical culture. Neurotoxicology.2009;30:47–51.

44. Vicente E, Boer M, Netto C, Fochesatto C, Dalmaz C, RodriguesSiqueira I, et al. Hippocampal antioxidant system in neonatesfrom methylmercury-intoxicated rats. Neurotoxicol Teratol.2004;26:817–23.

45. Sajdel-Sulkowska EM. Oxidative stress and neurotrophin signalingin autism. In: Chauhan A, Chauhan V, Brown WT, editors. Autism:oxidative stress, inflammation and immune abnormalities. BocaRaton, FL: CRC; 2010. p. 47–60.

46. Beal MF. Oxidatively modified proteins in aging and disease. FreeRadic Biol Med. 2002;32:797–803.

47. Neumann H, Hazen L, Weinstein J, Mehl RA, Chin JW.Genetically encoding protein oxidative damage. J Am ChemSoc. 2008;130:4028–33.

48. Gereben B, Zavacki AM, Ribich S, Kim BW, Huang SA, SimonidesWS, et al. Cellular and molecular basis of deiodinase-regulatedthyroid hormone signaling. Endocr Rev. 2008;29:898–938.

49. Galton VA, Wood ET, St Germain EA, Withrow CA, Aldrich G,St Germain GM, et al. Thyroid hormone homeostasis and actionin the type 2 deiodinase-deficient rodent brain during develop-ment. Endocrinology. 2007;148:3080.

50. Goldey ES, Kehn LS, Rehnberg GL, Crofton KM. Effects ofdevelopmental hypothyroidism on auditory and motor function inthe rat. Toxicol Appl Pharmacol. 1995;135:67–76.

51. Ng L, Goodyear RJ, Woods CA, Schneider MJ, Diamond E,Richardson GP, et al. Hearing loss and retarded cochleardevelopment in mice lacking type 2 iodothyronine deiodinase.Proc Natl Acad Sci U S A. 2004;101:3474–9.

52. Soldin OP, O’Mara DM, Aschner M. Thyroid hormones andmethylmercury toxicity. Biol Trace Elem Res. 2008;126:1–12.

53. Watanabe C. Selenium deficiency and brain functions: the significancefor methylmercury toxicity. Nippon Eiseigaku Zasshi. 2001;55:581–9.

54. Lamirand A, Pallud-Mothré S, Ramaugé M, Pierre M, Courtin F.Oxidative stress regulates type 3 deiodinase and type 2 deiodinasein cultured rat astrocytes. Endocrinology. 2008;149:3713–21.

55. Kim CY, Watanabe C, Satoh H. Effects of buthionine sulfoximine(BSO) on mercury distribution after Hg(o) exposure. Toxicology.1995;98:67–72.

56. James SJ, Slikker 3rd W, Melnyk S, New E, Pogribna M, Jernigan S.Thimerosal neurotoxicity is associated with glutathione deple-tion: protection with glutathione precursors. Neurotoxicology.2005;26:1–8.

57. Goswami A, Rosenberg I. Effects of glutathione on iodothyronine5′-deiodinase activity. Endocrinology. 1988;123:192–202.

58. CroteauW, Bodwell JE, Richardson JM, St Germain DL. Conservedcysteines in the type 1 deiodinase selenoprotein are not essential forcatalytic activity. J Biol Chem. 1998;273:25230–6.

59. Goemann IM, Gereben B, Harney JW, Zhu B, Maia AL, LarsenPR. Substitution of serine for proline in the active center of type 2iodothyronine deiodinase substantially alters its in vitro biochemicalproperties with dithiothreitol but not its function in intact cells.Endocrinology. 2010;151:821–9.

60. Dasgupta A, Das S, Sarkar PK. Thyroid hormone promotesglutathione synthesis in astrocytes by up regulation of glutamatecysteine ligase through differential stimulation of its catalytic andmodulator subunit mRNAs. Free Radic Biol Med. 2007;42:617–26.

61. Sobutskii MP, Kovan’ko EG, Liutinskii SI, Ivanov SD. Effect ofage and gender on genotoxic and biochemical indexes in animalblood after low doses of radiation-mercury exposure. AdvGerontol. 2007;20:91–6.

62. Olczak M, Duszczyk M, Mierzejewski P, Majewska MD.Neonatal administration of a vaccine preservative, thimerosal,produces lasting impairment of nociception and apparent activationof opioid system in rats. Brain Res. 2009;1301:143–51.

63. Holmes E, Nicholls AW, Lindon JC, Connor SC, Connelly JC,Haselden JN, et al. Chemometric models for toxicity classificationbased on NMR spectra of biofluids. Chem Res Toxicol.2000;13:471–8.

64. Gundacker C, Gencik M, Hengstschläger M. The relevance of theindividual genetic background for the toxicokinetics of twosignificant neurodevelopmental toxicants: mercury and lead.Mutat Res. 2010;705:130–40.

65. Ballerio R, Gianazza E, Mussoni L, Miller I, Gelosa P, Guerrini U,et al. Gender differences in endothelial function and inflammatorymarkers along the occurrence of pathological events in stroke-prone rats. Exp Mol Pathol. 2007;82:33–41.

66. Agnish ND, Keller KA. The rationale for culling of rodent litters.Fundam Appl Toxicol. 1997;38:2–6.

67. Palmer AK, Ulbrich BC. The cult of culling. Fundam ApplToxicol. 1997;38:7–22.

586 Cerebellum (2012) 11:575–586